Publicaciones científicas
2026
Marqués-Sánchez, Pilar; Bermejo-Martínez, David; Quiroga-Sánchez, Enedina; Calvo-Ayuso, Natalia; Martínez-Fernández, María Cristina; Benítez-Andrades, José Alberto
A Sexual Affiliation Network of Men Who Have Sex With Men Practicing Risk Sexual Behaviors and Chemsex: A Two-Mode Approach Artículo de revista
En: Public Health Nursing, vol. 43, no 2, pp. 246–255, 2026, ISSN: 1525-1446, (_eprint: https://onlinelibrary.wiley.com/doi/pdf/10.1111/phn.70058).
@article{marques-sanchez_sexual_2026,
title = {A Sexual Affiliation Network of Men Who Have Sex With Men Practicing Risk Sexual Behaviors and Chemsex: A Two-Mode Approach},
author = {Pilar Marqués-Sánchez and David Bermejo-Martínez and Enedina Quiroga-Sánchez and Natalia Calvo-Ayuso and María Cristina Martínez-Fernández and José Alberto Benítez-Andrades},
url = {https://onlinelibrary.wiley.com/doi/abs/10.1111/phn.70058},
doi = {10.1111/phn.70058},
issn = {1525-1446},
year = {2026},
date = {2026-01-01},
urldate = {2026-03-09},
journal = {Public Health Nursing},
volume = {43},
number = {2},
pages = {246–255},
abstract = {Background Dating apps for men who have sex with men (MSM) have facilitated unprotected sexual encounters and practices. In addition, drug use during such encounters is widespread among MSM. Traditionally, these populations have been studied in order to relate these facts to their socioeconomic characteristics and not from the perspective of the structure of their relationships, an aspect included in all nursing metaparadigms. Aim Describe the structure of the MSM affiliation network and their dating apps. Methods This was a descriptive cross-sectional study with 32 participants recruited online through the apps. The data used by this article come from the same sample and the same questionnaires as a previously conducted study; full details are provided in the Methods section. Results MSM demonstrated varying degrees, engaging with 1–5 applications (normalized degree: 0.200–0.494), while venue popularity spanned from 1 to 32 (normalized degree: 0.031–0.500). The core of the network, including two applications and 16 central MSM units, exhibited a higher density, that is, a high number of connections (0.593), compared to the periphery (0.050), indicating significant centralization. Conclusion The sexual affiliation network of MSM forms a cohesive, extensive network, with higher app usage affinity seen among individuals who use drugs, are from different birth countries, engage in group sex, or identify with nonhomosexual orientations.},
note = {_eprint: https://onlinelibrary.wiley.com/doi/pdf/10.1111/phn.70058},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Background Dating apps for men who have sex with men (MSM) have facilitated unprotected sexual encounters and practices. In addition, drug use during such encounters is widespread among MSM. Traditionally, these populations have been studied in order to relate these facts to their socioeconomic characteristics and not from the perspective of the structure of their relationships, an aspect included in all nursing metaparadigms. Aim Describe the structure of the MSM affiliation network and their dating apps. Methods This was a descriptive cross-sectional study with 32 participants recruited online through the apps. The data used by this article come from the same sample and the same questionnaires as a previously conducted study; full details are provided in the Methods section. Results MSM demonstrated varying degrees, engaging with 1–5 applications (normalized degree: 0.200–0.494), while venue popularity spanned from 1 to 32 (normalized degree: 0.031–0.500). The core of the network, including two applications and 16 central MSM units, exhibited a higher density, that is, a high number of connections (0.593), compared to the periphery (0.050), indicating significant centralization. Conclusion The sexual affiliation network of MSM forms a cohesive, extensive network, with higher app usage affinity seen among individuals who use drugs, are from different birth countries, engage in group sex, or identify with nonhomosexual orientations.
Castro, Ana González; Benítez-Andrades, José Alberto; Leirós-Rodríguez, Raquel
Accurate Fall Risk Prediction in Older Adults: Integrating Sensor and Clinical Data Through Machine Learning Proceedings Article
En: Fernández, Aurelio López; Rodríguez-González, Alejandro; Leirós-Rodríguez, Raquel; Miquel, Christian Mata; Suárez, Víctor Manuel González (Ed.): Artificial Intelligence in Biomedicine, pp. 609–616, Springer Nature Switzerland, Cham, 2026, ISBN: 978-3-032-10661-2.
@inproceedings{castro_accurate_2026,
title = {Accurate Fall Risk Prediction in Older Adults: Integrating Sensor and Clinical Data Through Machine Learning},
author = {Ana González Castro and José Alberto Benítez-Andrades and Raquel Leirós-Rodríguez},
editor = {Aurelio López Fernández and Alejandro Rodríguez-González and Raquel Leirós-Rodríguez and Christian Mata Miquel and Víctor Manuel González Suárez},
doi = {10.1007/978-3-032-10661-2_45},
isbn = {978-3-032-10661-2},
year = {2026},
date = {2026-01-01},
booktitle = {Artificial Intelligence in Biomedicine},
pages = {609–616},
publisher = {Springer Nature Switzerland},
address = {Cham},
abstract = {Accurate identification of older adults at high risk of falling is essential to prevent injuries and implement effective interventions. This study evaluated the performance of several machine learning models in predicting fall risk using both accelerometric data from wearable sensors and non-accelerometric data including demographic, functional, and clinical variables. A dataset comprising 146 older women was used to train and compare seven algorithms: random forest, XGBoost, AdaBoost, LightGBM, Bayesian ridge regression, decision trees, and support vector regression. Predictive accuracy was assessed using mean squared error, mean absolute error, and the coefficient of determination.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Accurate identification of older adults at high risk of falling is essential to prevent injuries and implement effective interventions. This study evaluated the performance of several machine learning models in predicting fall risk using both accelerometric data from wearable sensors and non-accelerometric data including demographic, functional, and clinical variables. A dataset comprising 146 older women was used to train and compare seven algorithms: random forest, XGBoost, AdaBoost, LightGBM, Bayesian ridge regression, decision trees, and support vector regression. Predictive accuracy was assessed using mean squared error, mean absolute error, and the coefficient of determination.
Rubio-Martín, Sergio; Crespo-Álvaro, Arturo; García-Ordás, María Teresa; Serrano-García, Antonio; Franch-Pato, Clara Margarita; Benítez-Andrades, José Alberto
Concept Normalization in Psychiatry: Comparing Embedding and Lexical Methods for Spanish Clinical Text Proceedings Article
En: Fernández, Aurelio López; Rodríguez-González, Alejandro; Leirós-Rodríguez, Raquel; Miquel, Christian Mata; Suárez, Víctor Manuel González (Ed.): Artificial Intelligence in Biomedicine, pp. 282–296, Springer Nature Switzerland, Cham, 2026, ISBN: 978-3-032-10661-2.
@inproceedings{rubio-martin_concept_2026,
title = {Concept Normalization in Psychiatry: Comparing Embedding and Lexical Methods for Spanish Clinical Text},
author = {Sergio Rubio-Martín and Arturo Crespo-Álvaro and María Teresa García-Ordás and Antonio Serrano-García and Clara Margarita Franch-Pato and José Alberto Benítez-Andrades},
editor = {Aurelio López Fernández and Alejandro Rodríguez-González and Raquel Leirós-Rodríguez and Christian Mata Miquel and Víctor Manuel González Suárez},
doi = {10.1007/978-3-032-10661-2_22},
isbn = {978-3-032-10661-2},
year = {2026},
date = {2026-01-01},
booktitle = {Artificial Intelligence in Biomedicine},
pages = {282–296},
publisher = {Springer Nature Switzerland},
address = {Cham},
abstract = {The automatic normalization of clinical entities is a critical step for enabling structured analysis of free-text medical records. This study proposes and evaluates four distinct retrieval strategies for normalizing entities extracted via Named Entity Recognition (NER) from Spanish psychiatric emergency notes, using Unified Medical Language System (UMLS) as the target terminology. A total of 768 annotated entities were mapped using MiniLM, Multilingual BERT, lexical matching, and the UMLS API. Results show that the API-based approach yields the best performance (Hit@3 = 56.8% and Hit@5 = 57.9%), effectively balancing accuracy and computational efficiency. Although embedding-based methods such as MiniLM and Multilingual BERT often outperform traditional techniques in other domains, they showed only marginal improvements over simple lexical matching in this context, while incurring significantly higher computational costs. These findings suggest that in the psychiatric domain, where language is often ambiguous, embedding-based approaches may fall short. Hybrid systems that combine fast retrieval with semantic reasoning are needed to capture the richness of psychiatric narratives.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The automatic normalization of clinical entities is a critical step for enabling structured analysis of free-text medical records. This study proposes and evaluates four distinct retrieval strategies for normalizing entities extracted via Named Entity Recognition (NER) from Spanish psychiatric emergency notes, using Unified Medical Language System (UMLS) as the target terminology. A total of 768 annotated entities were mapped using MiniLM, Multilingual BERT, lexical matching, and the UMLS API. Results show that the API-based approach yields the best performance (Hit@3 = 56.8% and Hit@5 = 57.9%), effectively balancing accuracy and computational efficiency. Although embedding-based methods such as MiniLM and Multilingual BERT often outperform traditional techniques in other domains, they showed only marginal improvements over simple lexical matching in this context, while incurring significantly higher computational costs. These findings suggest that in the psychiatric domain, where language is often ambiguous, embedding-based approaches may fall short. Hybrid systems that combine fast retrieval with semantic reasoning are needed to capture the richness of psychiatric narratives.
Benítez-Andrades, José Alberto; Laballos-González, Luis; Pujante-Otalora, Lorena; Nieto-Marcos, Sonia; Pinto-Carral, Arrate; Álvarez-Álvarez, María José
Machine learning-based prediction of disability in subacute low back pain: a primary care study on clinical and psychosocial determinants Artículo de revista
En: BMC Medical Informatics and Decision Making, vol. 26, no 1, pp. 92, 2026, ISSN: 1472-6947.
@article{benitez-andrades_machine_2026,
title = {Machine learning-based prediction of disability in subacute low back pain: a primary care study on clinical and psychosocial determinants},
author = {José Alberto Benítez-Andrades and Luis Laballos-González and Lorena Pujante-Otalora and Sonia Nieto-Marcos and Arrate Pinto-Carral and María José Álvarez-Álvarez},
url = {https://doi.org/10.1186/s12911-026-03384-6},
doi = {10.1186/s12911-026-03384-6},
issn = {1472-6947},
year = {2026},
date = {2026-02-01},
urldate = {2026-03-31},
journal = {BMC Medical Informatics and Decision Making},
volume = {26},
number = {1},
pages = {92},
abstract = {Subacute low back pain (LBP) is a highly prevalent condition and a major contributor to disability and health care burden. Early identification of individuals at risk of poor functional recovery is essential to support decision-making in primary care. Although prior research has identified relevant clinical and psychosocial predictors, the application of machine learning techniques for modeling disability and pain outcomes in this population remains limited.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Subacute low back pain (LBP) is a highly prevalent condition and a major contributor to disability and health care burden. Early identification of individuals at risk of poor functional recovery is essential to support decision-making in primary care. Although prior research has identified relevant clinical and psychosocial predictors, the application of machine learning techniques for modeling disability and pain outcomes in this population remains limited.
2025
Bayón-Gutiérrez, Martín; Prieto-Fernández, Natalia; García-Ordás, María Teresa; Benítez-Andrades, José Alberto; Alaiz-Moretón, Héctor; Grisetti, Giorgio
CAD2SLAM: Adaptive Projection Between CAD Blueprints and SLAM Maps Artículo de revista
En: IEEE Robotics and Automation Letters, vol. 10, no 2, pp. 1529–1536, 2025, ISSN: 2377-3766.
@article{bayon-gutierrez_cad2slam_2025,
title = {CAD2SLAM: Adaptive Projection Between CAD Blueprints and SLAM Maps},
author = {Martín Bayón-Gutiérrez and Natalia Prieto-Fernández and María Teresa García-Ordás and José Alberto Benítez-Andrades and Héctor Alaiz-Moretón and Giorgio Grisetti},
url = {https://ieeexplore.ieee.org/document/10816387},
doi = {10.1109/LRA.2024.3522838},
issn = {2377-3766},
year = {2025},
date = {2025-02-01},
urldate = {2025-08-30},
journal = {IEEE Robotics and Automation Letters},
volume = {10},
number = {2},
pages = {1529–1536},
abstract = {Robotic mobile platforms are key building blocks for numerous applications and cooperation between robots and humans is a key aspect to enhance productivity and reduce labor cost. To operate safely, robots typically rely on a custom map of the environment that depends on the sensor configuration of the platform. In contrast, blueprints stand as an abstract representation of the environment. The use of both CAD and SLAM maps allows robots to collaborate using the blueprint as a common language, while also easing the control for non-robotics experts. In this work we present an adaptive system to project a 2D pose in the blueprint to the SLAM map and vice-versa. Previous work in the literature aims at morphing a SLAM map to a previously available map. In contrast, CAD2SLAM does not alter the internal map representation used by the SLAM and localization algorithms running on the robot, preserving its original properties. We believe that our system is beneficial for the control and supervision of multiple heterogeneous robotic platforms that are monitored and controlled through the CAD map. Finally, we present a set of experiments that support our claims as well as open-source implementation.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Robotic mobile platforms are key building blocks for numerous applications and cooperation between robots and humans is a key aspect to enhance productivity and reduce labor cost. To operate safely, robots typically rely on a custom map of the environment that depends on the sensor configuration of the platform. In contrast, blueprints stand as an abstract representation of the environment. The use of both CAD and SLAM maps allows robots to collaborate using the blueprint as a common language, while also easing the control for non-robotics experts. In this work we present an adaptive system to project a 2D pose in the blueprint to the SLAM map and vice-versa. Previous work in the literature aims at morphing a SLAM map to a previously available map. In contrast, CAD2SLAM does not alter the internal map representation used by the SLAM and localization algorithms running on the robot, preserving its original properties. We believe that our system is beneficial for the control and supervision of multiple heterogeneous robotic platforms that are monitored and controlled through the CAD map. Finally, we present a set of experiments that support our claims as well as open-source implementation.
Pastor-Vargas, Rafael; Antón-Munárriz, Cristina; Haut, Juan M.; Robles-Gómez, Antonio; Paoletti, Mercedes E.; Benítez-Andrades, José Alberto
Cerebral ischemia detection using deep learning techniques Artículo de revista
En: Health Information Science and Systems, vol. 13, no 1, pp. 36, 2025, ISSN: 2047-2501.
@article{pastor-vargas_cerebral_2025,
title = {Cerebral ischemia detection using deep learning techniques},
author = {Rafael Pastor-Vargas and Cristina Antón-Munárriz and Juan M. Haut and Antonio Robles-Gómez and Mercedes E. Paoletti and José Alberto Benítez-Andrades},
url = {https://doi.org/10.1007/s13755-025-00352-8},
doi = {10.1007/s13755-025-00352-8},
issn = {2047-2501},
year = {2025},
date = {2025-05-01},
urldate = {2025-08-30},
journal = {Health Information Science and Systems},
volume = {13},
number = {1},
pages = {36},
abstract = {Cerebrovascular accident (CVA), commonly known as stroke, stands as a significant contributor to contemporary mortality and morbidity rates, often leading to lasting disabilities. Early identification is crucial in mitigating its impact and reducing mortality. Non-contrast computed tomography (NCCT) remains the primary diagnostic tool in stroke emergencies due to its speed, accessibility, and cost-effectiveness. NCCT enables the exclusion of hemorrhage and directs attention to ischemic causes resulting from arterial flow obstruction. Quantification of NCCT findings employs the Alberta Stroke Program Early Computed Tomography Score (ASPECTS), which evaluates affected brain structures. This study seeks to identify early alterations in NCCT density in patients with stroke symptoms using a binary classifier distinguishing NCCT scans with and without stroke. To achieve this, various well-known deep learning architectures, namely VGG3D, ResNet3D, and DenseNet3D, validated in the ImageNet challenges, are implemented with 3D images covering the entire brain volume. The training results of these networks are presented, wherein diverse parameters are examined for optimal performance. The DenseNet3D network emerges as the most effective model, attaining a training set accuracy of 98% and a test set accuracy of 95%. The aim is to alert medical professionals to potential stroke cases in their early stages based on NCCT findings displaying altered density patterns.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Cerebrovascular accident (CVA), commonly known as stroke, stands as a significant contributor to contemporary mortality and morbidity rates, often leading to lasting disabilities. Early identification is crucial in mitigating its impact and reducing mortality. Non-contrast computed tomography (NCCT) remains the primary diagnostic tool in stroke emergencies due to its speed, accessibility, and cost-effectiveness. NCCT enables the exclusion of hemorrhage and directs attention to ischemic causes resulting from arterial flow obstruction. Quantification of NCCT findings employs the Alberta Stroke Program Early Computed Tomography Score (ASPECTS), which evaluates affected brain structures. This study seeks to identify early alterations in NCCT density in patients with stroke symptoms using a binary classifier distinguishing NCCT scans with and without stroke. To achieve this, various well-known deep learning architectures, namely VGG3D, ResNet3D, and DenseNet3D, validated in the ImageNet challenges, are implemented with 3D images covering the entire brain volume. The training results of these networks are presented, wherein diverse parameters are examined for optimal performance. The DenseNet3D network emerges as the most effective model, attaining a training set accuracy of 98% and a test set accuracy of 95%. The aim is to alert medical professionals to potential stroke cases in their early stages based on NCCT findings displaying altered density patterns.
Albert, Elvira; Hähnle, Reiner; Merayo, Alicia; Steinhöfel, Dominic
Certified Cost Bounds for Abstract Programs Artículo de revista
En: ACM Trans. Softw. Eng. Methodol., vol. 34, no 3, pp. 67:1–67:33, 2025, ISSN: 1049-331X.
@article{albert_certified_2025,
title = {Certified Cost Bounds for Abstract Programs},
author = {Elvira Albert and Reiner Hähnle and Alicia Merayo and Dominic Steinhöfel},
url = {https://dl.acm.org/doi/10.1145/3705298},
doi = {10.1145/3705298},
issn = {1049-331X},
year = {2025},
date = {2025-02-01},
urldate = {2025-08-30},
journal = {ACM Trans. Softw. Eng. Methodol.},
volume = {34},
number = {3},
pages = {67:1–67:33},
abstract = {A program containing placeholders for unspecified statements or expressions is called an abstract (or schematic) program. Placeholder symbols occur naturally in program transformation rules, as used in refactoring, compilation or optimization. Static cost analysis derives the precise cost—or upper and lower bounds for it—of executing programs, as functions in terms of the program's input data size. We present a generalization of automated cost analysis that can handle abstract programs and, hence, can analyze the impact on the cost effect of program transformations. This kind of relational property requires provably precise cost bounds which are not always produced by cost analysis. Therefore, we certify by deductive verification that the inferred abstract cost bounds are correct and sufficiently precise. It is the first approach solving this problem. Both, abstract cost analysis and certification, are based on quantitative abstract execution (QAE) which in turn is a variation of abstract execution, a recently developed symbolic execution technique for abstract programs. To realize QAE the new concept of a cost invariant is introduced. QAE is implemented and runs fully automatically on a benchmark set consisting of representative optimization rules.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
A program containing placeholders for unspecified statements or expressions is called an abstract (or schematic) program. Placeholder symbols occur naturally in program transformation rules, as used in refactoring, compilation or optimization. Static cost analysis derives the precise cost—or upper and lower bounds for it—of executing programs, as functions in terms of the program's input data size. We present a generalization of automated cost analysis that can handle abstract programs and, hence, can analyze the impact on the cost effect of program transformations. This kind of relational property requires provably precise cost bounds which are not always produced by cost analysis. Therefore, we certify by deductive verification that the inferred abstract cost bounds are correct and sufficiently precise. It is the first approach solving this problem. Both, abstract cost analysis and certification, are based on quantitative abstract execution (QAE) which in turn is a variation of abstract execution, a recently developed symbolic execution technique for abstract programs. To realize QAE the new concept of a cost invariant is introduced. QAE is implemented and runs fully automatically on a benchmark set consisting of representative optimization rules.
Nforh, Lawrence; Michelena, Álvaro; Aveleira-Mata, Jose; García-Ordás, María Teresa; Zayas-Gato, Francisco; Jove, Esteban; Alaiz-Moretón, Héctor
DoS Attack Detection and Identification over Zigbee Environments Using Supervised Classification Algorithms Proceedings Article
En: Novais, Paulo; D., Parameshachari B.; Satoh, Ichiro; Inglada, Vicente Julian; González, Sara Rodríguez; Pérez, Esteban Jove; Domínguez, Javier Parra; Chamoso, Pablo; Alonso, Ricardo S. (Ed.): Ambient Intelligence – Software and Applications – 15th International Symposium on Ambient Intelligence, pp. 337–347, Springer Nature Switzerland, Cham, 2025, ISBN: 978-3-031-83117-1.
@inproceedings{nforh_dos_2025,
title = {DoS Attack Detection and Identification over Zigbee Environments Using Supervised Classification Algorithms},
author = {Lawrence Nforh and Álvaro Michelena and Jose Aveleira-Mata and María Teresa García-Ordás and Francisco Zayas-Gato and Esteban Jove and Héctor Alaiz-Moretón},
editor = {Paulo Novais and Parameshachari B. D. and Ichiro Satoh and Vicente Julian Inglada and Sara Rodríguez González and Esteban Jove Pérez and Javier Parra Domínguez and Pablo Chamoso and Ricardo S. Alonso},
doi = {10.1007/978-3-031-83117-1_32},
isbn = {978-3-031-83117-1},
year = {2025},
date = {2025-01-01},
booktitle = {Ambient Intelligence – Software and Applications – 15th International Symposium on Ambient Intelligence},
pages = {337–347},
publisher = {Springer Nature Switzerland},
address = {Cham},
abstract = {The Zigbee protocol, designed for low-power personal area wireless networks, is a technology widely used on the Internet of Things. This paper presents a study on the detection of denial-of-service attacks in Zigbee networks using supervised classification algorithms. Three techniques are evaluated: Logistic Regression, K-Nearest Neighbors and Support Vector Machines. A generated dataset is used for the analysis, and the results show that the K-Nearest Neighbors and Support Vector Machines approach achieves high performance and low computational demand. This methodology offers a promising strategy for security in Zigbee networks.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The Zigbee protocol, designed for low-power personal area wireless networks, is a technology widely used on the Internet of Things. This paper presents a study on the detection of denial-of-service attacks in Zigbee networks using supervised classification algorithms. Three techniques are evaluated: Logistic Regression, K-Nearest Neighbors and Support Vector Machines. A generated dataset is used for the analysis, and the results show that the K-Nearest Neighbors and Support Vector Machines approach achieves high performance and low computational demand. This methodology offers a promising strategy for security in Zigbee networks.
Benítez-Andrades, José Alberto; Prada-García, Camino; Ordás-Reyes, Nicolás; Blanco, Marta Esteban; Merayo, Alicia; Serrano-García, Antonio
Enhanced prediction of spine surgery outcomes using advanced machine learning techniques and oversampling methods Artículo de revista
En: Health Information Science and Systems, vol. 13, no 1, pp. 24, 2025, ISSN: 2047-2501.
@article{benitez-andrades_enhanced_2025,
title = {Enhanced prediction of spine surgery outcomes using advanced machine learning techniques and oversampling methods},
author = {José Alberto Benítez-Andrades and Camino Prada-García and Nicolás Ordás-Reyes and Marta Esteban Blanco and Alicia Merayo and Antonio Serrano-García},
url = {https://doi.org/10.1007/s13755-025-00343-9},
doi = {10.1007/s13755-025-00343-9},
issn = {2047-2501},
year = {2025},
date = {2025-03-01},
urldate = {2025-08-30},
journal = {Health Information Science and Systems},
volume = {13},
number = {1},
pages = {24},
abstract = {Accurate prediction of spine surgery outcomes is essential for optimizing treatment strategies. This study presents an enhanced machine learning approach to classify and predict the success of spine surgeries, incorporating advanced oversampling techniques and grid search optimization to improve model performance.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Accurate prediction of spine surgery outcomes is essential for optimizing treatment strategies. This study presents an enhanced machine learning approach to classify and predict the success of spine surgeries, incorporating advanced oversampling techniques and grid search optimization to improve model performance.
González-Castro, Ana; Benítez-Andrades, José Alberto; González-González, Rubén; Prada-García, Camino; Leirós-Rodríguez, Raquel
Predicting fall risk in older adults: A machine learning comparison of accelerometric and non-accelerometric factors Artículo de revista
En: DIGITAL HEALTH, vol. 11, pp. 20552076251331752, 2025, ISSN: 2055-2076, (Publisher: SAGE Publications Ltd).
@article{gonzalez-castro_predicting_2025,
title = {Predicting fall risk in older adults: A machine learning comparison of accelerometric and non-accelerometric factors},
author = {Ana González-Castro and José Alberto Benítez-Andrades and Rubén González-González and Camino Prada-García and Raquel Leirós-Rodríguez},
url = {https://journals.sagepub.com/action/showAbstract},
doi = {10.1177/20552076251331752},
issn = {2055-2076},
year = {2025},
date = {2025-03-01},
urldate = {2025-03-28},
journal = {DIGITAL HEALTH},
volume = {11},
pages = {20552076251331752},
publisher = {SAGE Publications Ltd},
abstract = {ObjectivesAccurate prediction of fall risk in older adults is essential to prevent injuries and improve quality of life. This study evaluates the predictive performance of various machine learning models using accelerometric data, non-accelerometric data, aiming to improve predictive accuracy and identify key contributing variable.MethodsWe applied random forest, XGBoost, AdaBoost, LightGBM, support vector regression (SVR), decision trees, and Bayesian ridge regression to a dataset of 146 older adults. Models were trained using accelerometric data (movement patterns) and non-accelerometric data (demographic and clinical variables). Performance was evaluated based on mean squared error (MSE) and coefficient of determination (R2), to assess how combining multiple data types influences prediction accuracy.ResultsModels trained on combined accelerometric and non-accelerometric data consistently outperformed those based on single data types. Bayesian ridge regression achieved the highest accuracy (MSE = 0.6746},
note = {Publisher: SAGE Publications Ltd},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
ObjectivesAccurate prediction of fall risk in older adults is essential to prevent injuries and improve quality of life. This study evaluates the predictive performance of various machine learning models using accelerometric data, non-accelerometric data, aiming to improve predictive accuracy and identify key contributing variable.MethodsWe applied random forest, XGBoost, AdaBoost, LightGBM, support vector regression (SVR), decision trees, and Bayesian ridge regression to a dataset of 146 older adults. Models were trained using accelerometric data (movement patterns) and non-accelerometric data (demographic and clinical variables). Performance was evaluated based on mean squared error (MSE) and coefficient of determination (R2), to assess how combining multiple data types influences prediction accuracy.ResultsModels trained on combined accelerometric and non-accelerometric data consistently outperformed those based on single data types. Bayesian ridge regression achieved the highest accuracy (MSE = 0.6746
Martínez-Villamea, Silvia; Prada-García, Camino; Benítez-Andrades, José Alberto; Quiroga-Sánchez, Enedina; García-Fernández, Rubén; Arias-Ramos, Natalia
Sleep Disturbances and Dietary Habits in Autism: A Comparative Analysis Artículo de revista
En: Journal of Autism and Developmental Disorders, 2025, ISSN: 1573-3432.
@article{martinez-villamea_sleep_2025,
title = {Sleep Disturbances and Dietary Habits in Autism: A Comparative Analysis},
author = {Silvia Martínez-Villamea and Camino Prada-García and José Alberto Benítez-Andrades and Enedina Quiroga-Sánchez and Rubén García-Fernández and Natalia Arias-Ramos},
url = {https://doi.org/10.1007/s10803-025-06964-z},
doi = {10.1007/s10803-025-06964-z},
issn = {1573-3432},
year = {2025},
date = {2025-07-01},
urldate = {2025-08-30},
journal = {Journal of Autism and Developmental Disorders},
abstract = {This study investigates dietary patterns and sleep quality in children and adolescents on the autism spectrum, compared to non-autistic peers. It explores the relationship between dietary habits and sleep quality, aiming to identify modifiable factors that could enhance well-being in ASD individuals. A cross-sectional case–control study was conducted with 158 participants on the autism spectrum and 77 non-autistic individuals aged 6–17 years in Spain. Dietary patterns were assessed using a validated food frequency questionnaire, while sleep quality was measured with the Children’s Sleep Habits Questionnaire (CSHQ-SP) and Pittsburgh Sleep Quality Index (PSQI). Statistical analyses, including non-parametric tests and Spearman’s correlation, were performed to examine differences and associations. Children on the autism spectrum displayed higher sugar intake and lower consumption of fruits and vegetables compared to non-autistic peers. ASD adolescents consumed more sugary beverages, with less pronounced differences in other food categories. Sleep quality was significantly poorer in the ASD group across all age cohorts, characterized by increased sleep latency, parasomnias, and daytime dysfunction. Positive associations were found between fruit and vegetable intake and improved sleep quality in ASD children. Unexpectedly, adolescents on the autism spectrum showed a complex relationship between sugar consumption and sleep quality, indicating potential short-term benefits but long-term risks. This study highlights the interplay between diet and sleep quality in ASD populations. Interventions promoting healthier eating habits, such as increased fruit and vegetable intake and reduced sugar consumption, could improve sleep outcomes and overall well-being in this vulnerable population.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
This study investigates dietary patterns and sleep quality in children and adolescents on the autism spectrum, compared to non-autistic peers. It explores the relationship between dietary habits and sleep quality, aiming to identify modifiable factors that could enhance well-being in ASD individuals. A cross-sectional case–control study was conducted with 158 participants on the autism spectrum and 77 non-autistic individuals aged 6–17 years in Spain. Dietary patterns were assessed using a validated food frequency questionnaire, while sleep quality was measured with the Children’s Sleep Habits Questionnaire (CSHQ-SP) and Pittsburgh Sleep Quality Index (PSQI). Statistical analyses, including non-parametric tests and Spearman’s correlation, were performed to examine differences and associations. Children on the autism spectrum displayed higher sugar intake and lower consumption of fruits and vegetables compared to non-autistic peers. ASD adolescents consumed more sugary beverages, with less pronounced differences in other food categories. Sleep quality was significantly poorer in the ASD group across all age cohorts, characterized by increased sleep latency, parasomnias, and daytime dysfunction. Positive associations were found between fruit and vegetable intake and improved sleep quality in ASD children. Unexpectedly, adolescents on the autism spectrum showed a complex relationship between sugar consumption and sleep quality, indicating potential short-term benefits but long-term risks. This study highlights the interplay between diet and sleep quality in ASD populations. Interventions promoting healthier eating habits, such as increased fruit and vegetable intake and reduced sugar consumption, could improve sleep outcomes and overall well-being in this vulnerable population.
Rubio-Martín, Sergio; García-Ordás, María Teresa; Corral-Fontecha, David; López-González, Laura; Alonso-Oláiz, Gonzalo; Crespo-©lvaro, Arturo; Benítez-Andrades, José Alberto
AI-Driven Survival Prediction in Pancreatic Cancer Proceedings Article
En: 2025 IEEE 38th International Symposium on Computer-Based Medical Systems (CBMS), pp. 284–289, 2025, ISSN: 2372-9198, (ISSN: 2372-9198).
@inproceedings{rubio-martin_ai-driven_2025,
title = {AI-Driven Survival Prediction in Pancreatic Cancer},
author = {Sergio Rubio-Martín and María Teresa García-Ordás and David Corral-Fontecha and Laura López-González and Gonzalo Alonso-Oláiz and Arturo Crespo-©lvaro and José Alberto Benítez-Andrades},
url = {https://ieeexplore.ieee.org/document/11058753},
doi = {10.1109/CBMS65348.2025.00064},
issn = {2372-9198},
year = {2025},
date = {2025-06-01},
urldate = {2025-07-13},
booktitle = {2025 IEEE 38th International Symposium on Computer-Based Medical Systems (CBMS)},
pages = {284–289},
abstract = {Pancreatic cancer remains one of the most aggressive malignancies, with limited survival rates and significant variability in patient outcomes. This study evaluates the performance of three machine learning models (Random Forest, Decision Tree, and XGBoost) in predicting patient survival at 3, 12, and 18 months, using data from the Complejo Asistencial Universitario de León (CAULE) Radiology Department. To systematically analyze the impact of different features on survival prediction, the dataset was structured into seven variable groups (G1G7), incorporating demographic, clinical, and treatment-related information. To address the inherent class imbalance in survival prediction, an Autoencoder-based synthetic data generation approach was applied, ensuring a balanced distribution of survival and non-survival cases across all timeframes. Hyperparameter tuning was performed, and experimental results indicate that Random Forest and XGBoost achieved comparable performance, both obtaining an accuracy above 81 % at 3 months, 83 % at 12 months, and 88 % at 18 months when trained on Group G7. To enhance model interpretability, SHapley Additive exPlanations (SHAP) was applied to the best-performing model, identifying key factors influencing survival.},
note = {ISSN: 2372-9198},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Pancreatic cancer remains one of the most aggressive malignancies, with limited survival rates and significant variability in patient outcomes. This study evaluates the performance of three machine learning models (Random Forest, Decision Tree, and XGBoost) in predicting patient survival at 3, 12, and 18 months, using data from the Complejo Asistencial Universitario de León (CAULE) Radiology Department. To systematically analyze the impact of different features on survival prediction, the dataset was structured into seven variable groups (G1G7), incorporating demographic, clinical, and treatment-related information. To address the inherent class imbalance in survival prediction, an Autoencoder-based synthetic data generation approach was applied, ensuring a balanced distribution of survival and non-survival cases across all timeframes. Hyperparameter tuning was performed, and experimental results indicate that Random Forest and XGBoost achieved comparable performance, both obtaining an accuracy above 81 % at 3 months, 83 % at 12 months, and 88 % at 18 months when trained on Group G7. To enhance model interpretability, SHapley Additive exPlanations (SHAP) was applied to the best-performing model, identifying key factors influencing survival.
Fontecha, David Corral; Fernández-Miranda, Pablo Menendez; Rubio-Martín, Sergio; Merayo-Corcoba, Alicia; González, Laura Lopez; Iglesias, Lara Lloret; Vega, Jose A
Enhancing Radiomic Feature Robustness Through Voxel Spacing-Aware Extraction in Anisotropic CT Data Proceedings Article
En: 2025 IEEE 38th International Symposium on Computer-Based Medical Systems (CBMS), pp. 490–495, 2025, (ISSN: 2372-9198).
@inproceedings{fontecha_enhancing_2025,
title = {Enhancing Radiomic Feature Robustness Through Voxel Spacing-Aware Extraction in Anisotropic CT Data},
author = {David Corral Fontecha and Pablo Menendez Fernández-Miranda and Sergio Rubio-Martín and Alicia Merayo-Corcoba and Laura Lopez González and Lara Lloret Iglesias and Jose A Vega},
url = {https://ieeexplore.ieee.org/document/11058871},
doi = {10.1109/CBMS65348.2025.00103},
year = {2025},
date = {2025-06-01},
urldate = {2025-08-30},
booktitle = {2025 IEEE 38th International Symposium on Computer-Based Medical Systems (CBMS)},
pages = {490–495},
abstract = {This study aimed to evaluate whether voxel spacing-aware radiomic feature extraction improves reproducibility, variability, and discriminative performance compared to conventional preprocessing methods in anisotropic CT data. A curated cohort of 685 pulmonary nodules from the LIDC-IDRI dataset was analyzed. Three preprocessing strategies-no resampling, isotropic resampling, and voxel spacing-aware extraction-and one postprocessing approach, voxel spacing weighting, were systematically compared. A modified version of PyRadiomics was developed to compute texture and shape features directly from native images while incorporating physical voxel dimensions without interpolation. Among the 94 extracted features, spacingaware preprocessing improved reproducibility in 58 features, reduced variability in 48, and enhanced univariate discrimination in 37. An ensemble feature selection combining six statistical and machine learning methods identified between 18 and 20 robust features per method. Logistic Regression models trained with spacing-aware features achieved the highest composite performance score (1.498), balancing discrimination and generalizability. SHAP interpretability analysis confirmed the clinical relevance of selected geometric and texture features. Overall, voxel spacing-aware preprocessing preserved native spatial structure and mitigated the effects of voxel geometry and interpolation artifacts, yielding stable and clinically interpretable radiomic features. These findings support the adoption of spacing-aware pipelines in heterogeneous and multi-center CT radiomics studies to enhance feature robustness and reproducibility.},
note = {ISSN: 2372-9198},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
This study aimed to evaluate whether voxel spacing-aware radiomic feature extraction improves reproducibility, variability, and discriminative performance compared to conventional preprocessing methods in anisotropic CT data. A curated cohort of 685 pulmonary nodules from the LIDC-IDRI dataset was analyzed. Three preprocessing strategies-no resampling, isotropic resampling, and voxel spacing-aware extraction-and one postprocessing approach, voxel spacing weighting, were systematically compared. A modified version of PyRadiomics was developed to compute texture and shape features directly from native images while incorporating physical voxel dimensions without interpolation. Among the 94 extracted features, spacingaware preprocessing improved reproducibility in 58 features, reduced variability in 48, and enhanced univariate discrimination in 37. An ensemble feature selection combining six statistical and machine learning methods identified between 18 and 20 robust features per method. Logistic Regression models trained with spacing-aware features achieved the highest composite performance score (1.498), balancing discrimination and generalizability. SHAP interpretability analysis confirmed the clinical relevance of selected geometric and texture features. Overall, voxel spacing-aware preprocessing preserved native spatial structure and mitigated the effects of voxel geometry and interpolation artifacts, yielding stable and clinically interpretable radiomic features. These findings support the adoption of spacing-aware pipelines in heterogeneous and multi-center CT radiomics studies to enhance feature robustness and reproducibility.
García-Ordás, María Teresa; Arcano-Bea, Paula; Rubiños, Manuel; Jove, Esteban; Narciandi-Rodriguez, Diego; Alaiz-Moretón, Héctor
Feature Importance Analysis of Meteorological Weather for Mini Eolic Electrical Power Prediction Using Clustering Information Proceedings Article
En: Quintián, Héctor; Corchado, Emilio; Lora, Alicia Troncoso; García, Hilde Pérez; Jove, Esteban; Rolle, José Luis Calvo; de Pisón, Francisco Javier Martínez; Bringas, Pablo García; Álvarez, Francisco Martínez; Cosío, Álvaro Herrero; Fosci, Paolo (Ed.): The 19th International Conference on Soft Computing Models in Industrial and Environmental Applications SOCO 2024, pp. 293–303, Springer Nature Switzerland, Cham, 2025, ISBN: 978-3-031-75010-6.
@inproceedings{garcia-ordas_feature_2025,
title = {Feature Importance Analysis of Meteorological Weather for Mini Eolic Electrical Power Prediction Using Clustering Information},
author = {María Teresa García-Ordás and Paula Arcano-Bea and Manuel Rubiños and Esteban Jove and Diego Narciandi-Rodriguez and Héctor Alaiz-Moretón},
editor = {Héctor Quintián and Emilio Corchado and Alicia Troncoso Lora and Hilde Pérez García and Esteban Jove and José Luis Calvo Rolle and Francisco Javier Martínez de Pisón and Pablo García Bringas and Francisco Martínez Álvarez and Álvaro Herrero Cosío and Paolo Fosci},
doi = {10.1007/978-3-031-75010-6_29},
isbn = {978-3-031-75010-6},
year = {2025},
date = {2025-01-01},
booktitle = {The 19th International Conference on Soft Computing Models in Industrial and Environmental Applications SOCO 2024},
pages = {293–303},
publisher = {Springer Nature Switzerland},
address = {Cham},
abstract = {The increasing concern about climate change has contributed to promoting renewable energy technologies. Mini-eolic turbines are a common solution for domestic energy supply self-consumption installations. Due to this technology’s strong dependency on climate conditions and corresponding variability, it is important to develop intelligent systems to model and estimate its behavior. This paper uses three different feature selection methods, a clustering algorithm, and a regression technique to predict the power generated by a small wind turbine located in a bioclimatic house. Different configurations are tested, evaluating the impact on the model performance.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The increasing concern about climate change has contributed to promoting renewable energy technologies. Mini-eolic turbines are a common solution for domestic energy supply self-consumption installations. Due to this technology’s strong dependency on climate conditions and corresponding variability, it is important to develop intelligent systems to model and estimate its behavior. This paper uses three different feature selection methods, a clustering algorithm, and a regression technique to predict the power generated by a small wind turbine located in a bioclimatic house. Different configurations are tested, evaluating the impact on the model performance.
Rubio-Martín, Sergio; Crespo-Álvaro, Arturo; García-Ordás, María Teresa; Serrano-García, Antonio; Franch-Pato, Clara Margarita; Benítez-Andrades, José Alberto
Fine-Tuning Transformer Models for Structuring Spanish Psychiatric Clinical Notes Proceedings Article
En: 2025 IEEE 38th International Symposium on Computer-Based Medical Systems (CBMS), pp. 97–102, 2025, (ISSN: 2372-9198).
@inproceedings{rubio-martin_fine-tuning_2025,
title = {Fine-Tuning Transformer Models for Structuring Spanish Psychiatric Clinical Notes},
author = {Sergio Rubio-Martín and Arturo Crespo-Álvaro and María Teresa García-Ordás and Antonio Serrano-García and Clara Margarita Franch-Pato and José Alberto Benítez-Andrades},
url = {https://ieeexplore.ieee.org/document/11058869},
doi = {10.1109/CBMS65348.2025.00028},
year = {2025},
date = {2025-06-01},
urldate = {2025-08-30},
booktitle = {2025 IEEE 38th International Symposium on Computer-Based Medical Systems (CBMS)},
pages = {97–102},
abstract = {The unstructured nature of psychiatric clinical notes poses a significant challenge for automated information extraction and data structuring. In this study, we explore the use of transformer-based language models to perform Named Entity Recognition (NER) on de-identified Spanish electronic health records (EHRs) provided by the Psychiatry Service of Complejo Asistencial Universitario de León (CAULE). A manually annotated gold standard, consisting of 200 clinical notes, was developed by domain experts to evaluate the performance of five models: BETO (cased and uncased), ALBETO, ClinicalBERT, and Bio_ClinicalBERT. Each model was fine-tuned and assessed using a strict exact matching criterion across six clinically relevant label types. Results demonstrate that ClinicalBERT, despite being pre-trained on English medical corpora, achieved the highest macro-average F1-score on the test set (80 %). However, BETO-cased outperformed ClinicalBERT in four out of six label types, being better in categories with higher syntactic variability. Lower-performing models, such as ALBETO and Bio_ClinicalBERT, struggled to generalize to Spanish psychiatric language, likely due to domain and language mismatches. This work highlights the effectiveness of transformer-based architectures for structuring psychiatric narratives in Spanish and provides a robust foundation for future clinical NLP applications in non-English contexts.},
note = {ISSN: 2372-9198},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The unstructured nature of psychiatric clinical notes poses a significant challenge for automated information extraction and data structuring. In this study, we explore the use of transformer-based language models to perform Named Entity Recognition (NER) on de-identified Spanish electronic health records (EHRs) provided by the Psychiatry Service of Complejo Asistencial Universitario de León (CAULE). A manually annotated gold standard, consisting of 200 clinical notes, was developed by domain experts to evaluate the performance of five models: BETO (cased and uncased), ALBETO, ClinicalBERT, and Bio_ClinicalBERT. Each model was fine-tuned and assessed using a strict exact matching criterion across six clinically relevant label types. Results demonstrate that ClinicalBERT, despite being pre-trained on English medical corpora, achieved the highest macro-average F1-score on the test set (80 %). However, BETO-cased outperformed ClinicalBERT in four out of six label types, being better in categories with higher syntactic variability. Lower-performing models, such as ALBETO and Bio_ClinicalBERT, struggled to generalize to Spanish psychiatric language, likely due to domain and language mismatches. This work highlights the effectiveness of transformer-based architectures for structuring psychiatric narratives in Spanish and provides a robust foundation for future clinical NLP applications in non-English contexts.
García-Ordás, María Teresa; Díaz-Longueira, Antonio; Michelena, Álvaro; Jove, Esteban; Bayón-Gutiérrez, Martín; Alaiz-Moretón, Héctor
Missing Meteorological Data Imputation for Mini Eolic Electrical Power Prediction Proceedings Article
En: Quintián, Héctor; Corchado, Emilio; Lora, Alicia Troncoso; García, Hilde Pérez; Pérez, Esteban Jove; Rolle, José Luis Calvo; de Pisón, Francisco Javier Martínez; Bringas, Pablo García; Álvarez, Francisco Martínez; Herrero, Álvaro; Fosci, Paolo (Ed.): Hybrid Artificial Intelligent Systems, pp. 78–87, Springer Nature Switzerland, Cham, 2025, ISBN: 978-3-031-74186-9.
@inproceedings{garcia-ordas_missing_2025,
title = {Missing Meteorological Data Imputation for Mini Eolic Electrical Power Prediction},
author = {María Teresa García-Ordás and Antonio Díaz-Longueira and Álvaro Michelena and Esteban Jove and Martín Bayón-Gutiérrez and Héctor Alaiz-Moretón},
editor = {Héctor Quintián and Emilio Corchado and Alicia Troncoso Lora and Hilde Pérez García and Esteban Jove Pérez and José Luis Calvo Rolle and Francisco Javier Martínez de Pisón and Pablo García Bringas and Francisco Martínez Álvarez and Álvaro Herrero and Paolo Fosci},
doi = {10.1007/978-3-031-74186-9_7},
isbn = {978-3-031-74186-9},
year = {2025},
date = {2025-01-01},
booktitle = {Hybrid Artificial Intelligent Systems},
pages = {78–87},
publisher = {Springer Nature Switzerland},
address = {Cham},
abstract = {The continuous rising trend shown by greenhouse emissions has led to a global situation in which the promotion of clean alternative technologies is crucial. In this context, small green power self-consumption installations represent an effective and clean solution to reduce climate change. However, they must be subjected to exhaustive supervision of the process, from mechanical, electrical, or electronic components, to ensure good performance and economic feasibility. This work proposes different data imputation techniques to deal with missing data derived from sensor missreadings in a minieolic installation. The performance of regression techniques over each reconstructed set is evaluated with successful results.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The continuous rising trend shown by greenhouse emissions has led to a global situation in which the promotion of clean alternative technologies is crucial. In this context, small green power self-consumption installations represent an effective and clean solution to reduce climate change. However, they must be subjected to exhaustive supervision of the process, from mechanical, electrical, or electronic components, to ensure good performance and economic feasibility. This work proposes different data imputation techniques to deal with missing data derived from sensor missreadings in a minieolic installation. The performance of regression techniques over each reconstructed set is evaluated with successful results.
Rubio-Martín, Sergio; García-Ordás, María Teresa; Serrano-García, Antonio; Franch-Pato, Clara Margarita; Crespo-Álvaro, Arturo; Benítez-Andrades, José Alberto
Classification of psychiatry clinical notes by diagnosis: a deep learning and machine learning approach Artículo de revista
En: PeerJ Computer Science, vol. 11, pp. e3045, 2025, ISSN: 2376-5992, (Publisher: PeerJ Inc.).
@article{rubio-martin_classification_2025,
title = {Classification of psychiatry clinical notes by diagnosis: a deep learning and machine learning approach},
author = {Sergio Rubio-Martín and María Teresa García-Ordás and Antonio Serrano-García and Clara Margarita Franch-Pato and Arturo Crespo-Álvaro and José Alberto Benítez-Andrades},
url = {https://peerj.com/articles/cs-3045},
doi = {10.7717/peerj-cs.3045},
issn = {2376-5992},
year = {2025},
date = {2025-07-01},
urldate = {2025-08-30},
journal = {PeerJ Computer Science},
volume = {11},
pages = {e3045},
abstract = {The classification of clinical notes into specific diagnostic categories is critical in healthcare, especially for mental health conditions like anxiety and adjustment disorder. In this study, we compare the performance of various artificial intelligence models, including both traditional machine learning approaches (random forest, support vector machine, K-nearest neighbors, decision tree, and eXtreme Gradient Boost) and deep learning models (DistilBERT and SciBERT), to classify clinical notes into these two diagnoses. Additionally, we implemented three oversampling strategies: No Oversampling, Random Oversampling, and Synthetic Minority Over-sampling Technique (SMOTE), to assess their impact on model performance. Hyperparameter tuning was also applied to optimize model accuracy. Our results indicate that oversampling techniques had minimal impact on model performance overall. The only exception was SMOTE, which showed a positive effect specifically with Bidirectional Encoder Representations from Transformers (BERT)-based models. However, hyperparameter optimization significantly improved accuracy across the models, enhancing their ability to generalize and perform on the dataset. The decision tree and eXtreme Gradient Boost models achieved the highest accuracy among machine learning approaches, both reaching 96%, while the DistilBERT and SciBERT models also attained 96% accuracy in the deep learning category. These findings underscore the importance of hyperparameter tuning in maximizing model performance. This study contributes to the ongoing research on AI-assisted diagnostic tools in mental health by providing insights into the efficacy of different model architectures and data balancing methods.},
note = {Publisher: PeerJ Inc.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
The classification of clinical notes into specific diagnostic categories is critical in healthcare, especially for mental health conditions like anxiety and adjustment disorder. In this study, we compare the performance of various artificial intelligence models, including both traditional machine learning approaches (random forest, support vector machine, K-nearest neighbors, decision tree, and eXtreme Gradient Boost) and deep learning models (DistilBERT and SciBERT), to classify clinical notes into these two diagnoses. Additionally, we implemented three oversampling strategies: No Oversampling, Random Oversampling, and Synthetic Minority Over-sampling Technique (SMOTE), to assess their impact on model performance. Hyperparameter tuning was also applied to optimize model accuracy. Our results indicate that oversampling techniques had minimal impact on model performance overall. The only exception was SMOTE, which showed a positive effect specifically with Bidirectional Encoder Representations from Transformers (BERT)-based models. However, hyperparameter optimization significantly improved accuracy across the models, enhancing their ability to generalize and perform on the dataset. The decision tree and eXtreme Gradient Boost models achieved the highest accuracy among machine learning approaches, both reaching 96%, while the DistilBERT and SciBERT models also attained 96% accuracy in the deep learning category. These findings underscore the importance of hyperparameter tuning in maximizing model performance. This study contributes to the ongoing research on AI-assisted diagnostic tools in mental health by providing insights into the efficacy of different model architectures and data balancing methods.
Benítez-Andrades, José Alberto; González-Castro, Ana; González-Marcos, Elena; Rubio-Martín, Sergio; Leirós-Rodríguez, Raquel; Prada-García, Camino
Feasibility of accelerometer-based prediction of postural balance in schoolchildren using machine learning models Artículo de revista
En: Scientific Reports, vol. 15, no 1, pp. 45349, 2025, ISSN: 2045-2322.
@article{benitez-andrades_feasibility_2025,
title = {Feasibility of accelerometer-based prediction of postural balance in schoolchildren using machine learning models},
author = {José Alberto Benítez-Andrades and Ana González-Castro and Elena González-Marcos and Sergio Rubio-Martín and Raquel Leirós-Rodríguez and Camino Prada-García},
url = {https://www.nature.com/articles/s41598-025-30160-9},
doi = {10.1038/s41598-025-30160-9},
issn = {2045-2322},
year = {2025},
date = {2025-11-01},
urldate = {2026-01-02},
journal = {Scientific Reports},
volume = {15},
number = {1},
pages = {45349},
publisher = {Nature Publishing Group},
abstract = {Objective assessment of balance during childhood is essential for supporting motor development, preventing falls, and identifying potential impairments. Traditional clinical tests, such as the Flamingo and balance beam, are widely used but remain limited by subjectivity and lack of precision. Accelerometry offers quantitative measures, and its integration with machine learning models provides an opportunity to explore predictive assessment in pediatric populations. This study evaluated the feasibility of predicting clinical balance test outcomes in schoolchildren using accelerometric data from both static and dynamic tasks. A cross-sectional study was conducted with 90 children aged 6 to 12 years. Accelerometric signals were recorded in three axes and magnitude during static tasks (eyes closed, eyes open, and unstable surface) and during gait. Outcomes included the number of supports in the Flamingo test and the distance covered in the balance beam test. Several machine learning models, including linear regression, penalized regression, k-nearest neighbors, support vector regression, and random forest, were applied using 5-fold cross-validation. Models showed modest but consistent predictive accuracy for the Flamingo test, particularly with static tasks, with random forest, support vector regression, and k-nearest neighbors performing best. Prediction of balance beam outcomes was poor across all models. These findings suggest that accelerometry-based machine learning is feasible for predicting balance performance in children, especially for the Flamingo test, supporting its potential as a digital tool for screening and educational health applications.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Objective assessment of balance during childhood is essential for supporting motor development, preventing falls, and identifying potential impairments. Traditional clinical tests, such as the Flamingo and balance beam, are widely used but remain limited by subjectivity and lack of precision. Accelerometry offers quantitative measures, and its integration with machine learning models provides an opportunity to explore predictive assessment in pediatric populations. This study evaluated the feasibility of predicting clinical balance test outcomes in schoolchildren using accelerometric data from both static and dynamic tasks. A cross-sectional study was conducted with 90 children aged 6 to 12 years. Accelerometric signals were recorded in three axes and magnitude during static tasks (eyes closed, eyes open, and unstable surface) and during gait. Outcomes included the number of supports in the Flamingo test and the distance covered in the balance beam test. Several machine learning models, including linear regression, penalized regression, k-nearest neighbors, support vector regression, and random forest, were applied using 5-fold cross-validation. Models showed modest but consistent predictive accuracy for the Flamingo test, particularly with static tasks, with random forest, support vector regression, and k-nearest neighbors performing best. Prediction of balance beam outcomes was poor across all models. These findings suggest that accelerometry-based machine learning is feasible for predicting balance performance in children, especially for the Flamingo test, supporting its potential as a digital tool for screening and educational health applications.
González-Castro, Ana; Leirós-Rodríguez, Raquel; Nistal-Martínez, Marta; Bodero-Vidal, Ernesto; Benítez-Andrades, José Alberto; Hernandez-Lucas, Pablo
Effect of COVID-19 on Falls in a Residential Care Facility for the Elderly: Longitudinal Observational Study Artículo de revista
En: Journal of Clinical Medicine, vol. 14, no 17, pp. 6229, 2025, ISSN: 2077-0383.
@article{gonzalez-castro_effect_2025,
title = {Effect of COVID-19 on Falls in a Residential Care Facility for the Elderly: Longitudinal Observational Study},
author = {Ana González-Castro and Raquel Leirós-Rodríguez and Marta Nistal-Martínez and Ernesto Bodero-Vidal and José Alberto Benítez-Andrades and Pablo Hernandez-Lucas},
url = {https://www.mdpi.com/2077-0383/14/17/6229},
doi = {10.3390/jcm14176229},
issn = {2077-0383},
year = {2025},
date = {2025-01-01},
urldate = {2026-01-02},
journal = {Journal of Clinical Medicine},
volume = {14},
number = {17},
pages = {6229},
publisher = {Multidisciplinary Digital Publishing Institute},
abstract = {Background/Objectives: During the Coronavirus Disease 2019 (COVID-19) pandemic, various safety measures were implemented in elderly care facilities in Spain. These measures led to a reduction in physical activity and increased supervision of residents, often resulting in the suspension of outings from the facility. The objective of this study was to assess the influence of COVID-19 preventive measures on the number and characteristics of falls among elderly individuals living in a residential care facility in Spain. Methods: A retrospective longitudinal observational study was conducted from 2018 to 2021. Over these four years, data related to falls were collected from a residential care facility for the elderly. Both patient characteristics and fall characteristics were recorded. Results: The average age of the 48 residents continuously institutionalized between 2018 and 2021 was 85.8 ± 5.1 years. A total of 364 falls occurred during the study period, with 68% of them taking place in 2019 and 2020. Although the number of falls increased during the COVID-19 pandemic, the characteristics of the falls did not change. However, residents who experienced falls were increasingly accompanied at the time of the event. Conclusions: Based on the data collected from the elderly care facility analyzed in this study, falls increased during the COVID-19 pandemic, but the measures implemented in residential care facilities do not appear to have altered the characteristics of the falls, except for the greater presence of companionship.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Background/Objectives: During the Coronavirus Disease 2019 (COVID-19) pandemic, various safety measures were implemented in elderly care facilities in Spain. These measures led to a reduction in physical activity and increased supervision of residents, often resulting in the suspension of outings from the facility. The objective of this study was to assess the influence of COVID-19 preventive measures on the number and characteristics of falls among elderly individuals living in a residential care facility in Spain. Methods: A retrospective longitudinal observational study was conducted from 2018 to 2021. Over these four years, data related to falls were collected from a residential care facility for the elderly. Both patient characteristics and fall characteristics were recorded. Results: The average age of the 48 residents continuously institutionalized between 2018 and 2021 was 85.8 ± 5.1 years. A total of 364 falls occurred during the study period, with 68% of them taking place in 2019 and 2020. Although the number of falls increased during the COVID-19 pandemic, the characteristics of the falls did not change. However, residents who experienced falls were increasingly accompanied at the time of the event. Conclusions: Based on the data collected from the elderly care facility analyzed in this study, falls increased during the COVID-19 pandemic, but the measures implemented in residential care facilities do not appear to have altered the characteristics of the falls, except for the greater presence of companionship.
Gutiérrez, Martín Bayón; Fernández, Natalia Prieto; Ordás, María Teresa García; Benavides, Carmen; Andrades, José Alberto Benítez
Supervisión de sistemas multi-robot mediante proyección adaptativa sobre mapa común Artículo de revista
En: Jornadas de Automática, no 46, 2025, ISSN: 3045-4093.
@article{gutierrez_supervision_2025,
title = {Supervisión de sistemas multi-robot mediante proyección adaptativa sobre mapa común},
author = {Martín Bayón Gutiérrez and Natalia Prieto Fernández and María Teresa García Ordás and Carmen Benavides and José Alberto Benítez Andrades},
url = {https://revistas.udc.es/index.php/JA_CEA/article/view/12270},
doi = {10.17979/ja-cea.2025.46.12270},
issn = {3045-4093},
year = {2025},
date = {2025-09-01},
urldate = {2026-03-31},
journal = {Jornadas de Automática},
number = {46},
abstract = {This manuscript presents a system for the adaptive representation of the pose of multiple robotic platforms operating in the same environment. The system is based on the CAD2SLAM methodology and projects the individual positions of each robot onto a common base map employing a non-rigid projection grid adjusted using a set of correspondence points and least-squares optimization. This enables a coherent and easy to understand representation of each robot’s relative position, regardless of the type of sensors or SLAM system. The human operator can supervise the global state of the multi-robot system in a more intuitive and efficient manner, without altering the internal architecture and the normal operation of the robots. Validated in both simulated and real-world scenarios, it demonstrates robustness and real-time execution. The main novelty of this system lies in the use of a base map as a common reference system for all robots, facilitating cooperation between heterogeneous platforms and improving the visual interpretation of the environment.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
This manuscript presents a system for the adaptive representation of the pose of multiple robotic platforms operating in the same environment. The system is based on the CAD2SLAM methodology and projects the individual positions of each robot onto a common base map employing a non-rigid projection grid adjusted using a set of correspondence points and least-squares optimization. This enables a coherent and easy to understand representation of each robot’s relative position, regardless of the type of sensors or SLAM system. The human operator can supervise the global state of the multi-robot system in a more intuitive and efficient manner, without altering the internal architecture and the normal operation of the robots. Validated in both simulated and real-world scenarios, it demonstrates robustness and real-time execution. The main novelty of this system lies in the use of a base map as a common reference system for all robots, facilitating cooperation between heterogeneous platforms and improving the visual interpretation of the environment.
